ENHANCED TRANSVERSE MIXING OF POLLUTANTS IN STREAM WITH SUBMERGED VANES

Abstract

Streams have been used for the disposal of various industrial and municipal wastes since long time. Quantitative understanding of mixing of such pollutants in streams is a matter of concern in recent years for the effective control of pollution in the streams. Most of the natural streams are relatively shallow compared with their length and width. When pollutants were disposed off at a point/section of a stream, it mixes quickly over the entire depth and then continues to spread in the longitudinal and transverse flow directions. Thus, it is essential to study how the waste/effluent gets mixed in the flowing stream for the environmental concern and water quality modeling. Excluding the initial distance required to achieve mixing in the vertical direction, the mixing of pollutants can be efficiently modeled by two-dimensional depthaveraged mixing equation, i.e., transverse mixing equation. Transverse mixing is arguably more important in water quality management than either vertical or longitudinal mixing, especially when dealing with the discharge of pollutants from point sources or the mixing of tributary inflows. In such problems, vertical mixing occurs rapidly and is only important very close to the source, whereas, longitudinal dispersion is only important in far-field if the source is unsteady. In straight open channels, secondary currents are weak compared to curved channels, therefore, spreading of pollutant is higher in curved channels compared to the straight channels (Krishnappan and Lau, 1981; Holley and Nerat, 1983; Boxall et al., 2003; Boxall and Guymer, 2003; Albers and Steffler, 2007; Dow et al., 2009). To increase the secondary current for the enhancement of transverse mixing, it is desirable to have some structure, which can increase the secondary current in that reach. Submerged vanes are suitable structure for this purpose. Transverse mixing has extensively been studied in straight, curved, meandering channels. However, no study related to the effect of artificially-induced secondary current on the transverse mixing has been conducted, so far. The present proposal is intended to study this aspect of the transverse mixing. Submerged vane is basically an aerofoil structure, which generates the excess turbulence in form of helical flow structure in the flow due to pressure difference between approaching flow side and downstream side of vane (Odgaard and Spoljaric, 1986; Odgaard and Mosconi, 1987; Odgaard and Wang, 1991; Wang and Odgaard, 1993). These vanes are in general placed at a certain angle with respect to the flow directions which is usually equal to 10o – 40o. Submerged vanes utilize vorticity to minimize the drag and produce flow redistribution in the flow such iii that longitudinal flow is compelled to get diverted towards the transverse direction (Wang and Odgaard, 1993). Many investigators like Odgaard and Wang (1991a), Wang and Odgaard (1993), Marelius and Sinha (1998), Tan et al. (2005), Ouyang et al. (2008) have studied analytically and experimentally the flow structure around the submerged vanes. Following were the objectives of the present study: (i) Development of a numerical scheme for the solution of unsteady transverse mixing equation. (ii) To study the secondary current induced by installation of series of submerged vanes on the channel beds. To optimize vane size, location, spacing and alignment for obtaining strong secondary current. (iii) Measurement of concentration profile of tracer injected either side of the channel in the presence of installed vanes and also without it for various flow conditions and vane configurations. (iv) Determination of mixing coefficients using the measured concentration profiles. Also development of a predictor for enhanced mixing coefficients incorporating the flow parameters, vane sizes and spacing. (v) Final recommendations for using submerged vanes for the enhancement of transverse mixing to be made. The governing mass balance equation of transient transverse mixing in streams has been solved using finite volume method invoking the weighted upwind scheme. The developed model takes care of variation of depth of flow, depth-averaged velocity and transverse mixing coefficient across the channel width. The developed model is validated with analytical and experimental data. The numerical model proposed by Ahmad (2008) is extended to determine transverse mixing coefficient from the known concentration profiles at the downstream stations. Satisfactory agreement is found between concentration profiles computed using the proposed finite volume model and the analytical model for constant mixing coefficient and continuous pollutant injection. Experiments were performed in a recirculating concrete flume of width 1.0 m, depth 0.30 m and length 19 m. The bed slope of the flume was 0.000632. The Rhodamine WT was used as iv tracer due to its high detectability and conservative in nature. A tracer injecting system was used to inject dye from one side of the flume that represents a plane source of width 100 mm. The Rhodamine WT dye concentration was measured across the width of the channel and downstream of injection point using Hydro lab MS-5 probe. Micro ADV was used to measure three-dimensional velocity field downstream of the injection location. Three sizes of vanes were used in the experimentations which were 0.02 m  0.05 m, 0.04 m  0.1 m and 0.06 m  0.12 m, whose lateral spacing was 0.05 m, 0.1 m and 0.125 m, respectively. Three flow conditions were maintained in order to perform experimentations for depth of flow of 0.09 cm, 0.1025 m and 0.1241 m. A total 50 runs were taken. Experiments were performed under no, one, two, three and four arrays of vanes for the measurement of dye concentration and threedimensional velocities. A computational fluid dynamics (CFD) model was developed on ANSYS-CFX platform to simulate flow pattern and turbulence characteristics around and downstream of a submerged vane and a series of vane rows containing multiple vanes in a single row. k- turbulence model is used herein. The developed CFD model is validated for the single vane, from the transverse velocity profile measured by Wang and Odgaard (1993) at x = 2H, 8H and 20H for vane size 0.076 m  0.152 m for depth of flow = 0.152 m. It was observed that for each of transect, the simulated transverse velocity profile matched with the observed transverse velocity profile in a satisfactory manner. It was observed that vorticity has got its maximum value when the angle of attack was 28.7o which was in accordance with the available literature. It was also observed that optimum value of height of vane which induced maximum intensity vortices was 0.4 times depth of flow. It was observed that as the length of vane was increased, the intensity of vorticity also increased subsequently. Simulated flow downstream of a vane indicates that the leading edge of the vane induces high vorticity which decreases exponentially downstream of the vane. The variation of turbulent kinetic energy is also on the pattern of vorticity. For the multiple vane arrays system, the developed CFD model is validated by comparing simulated longitudinal velocity at three sections viz. x = 3H, 8H and 20H with the measured values and validation was also done for measured transverse velocity with the simulated transverse velocity profile. Good agreement in the simulated and observed velocity profiles were observed. Simulated flow pattern around multiple array of vanes indicate that near to the submerged vanes a large vortical field exists. It was observed that when the lateral spacing of vanes in an array was kept at y = 3H (where H = Height of vane), maximum intensity of vorticity was induced in the flow. In order to optimize longitudinal spacing a method was v proposed in which to acquire vorticity of a given strength downstream of a pre-installed vane row, the distance is read from the calibrated graph in lieu of the % vorticity required from the vane row in downstream. In case of multiple vanes, it was observed that at a distance very near to vane row, each vane row generated a circulation of its own. In the further downstream, vortices coalesce with each other to form a larger field of circulation but effective magnitude of circulation was observed to be less than the magnitude of circulation if it was generated by each vane in individual manner. Moving further downstream, it was observed that circulation field was dissipated and less disturbed flow field was obtained under the action of viscosity. Turbulent kinetic energy was observed to be higher in the case of four arrays of vanes than the one array of vane. Analysis of measured instantaneous velocity downstream of the zero, one, two, three and four arrays of submerged vanes indicates that in the presence of vanes, flow near to the vane is highly unstable and chaotic. The turbulence is clearly having heterogeneity as going up in vertical direction. The turbulence quantities decrease from bed to flow surface. The variation of all turbulence characteristics was same in all directions and was nearly overlapping to each other for measured instantaneous velocity at three transects in case of plane shear flow. Variation of transverse velocity along depth for zero one, two, three and four arrays of vanes is also being studied. It is found that as array of vanes increases the transverse velocity increases which signify that transverse mixing shall be higher for higher number of arrays of vanes. Effect of submerged vanes on tracer concentration profiles was studied experimentally. For this purpose, variation of tracer concentration across the width of the channel at distance of 5 m and 15 m for the depth of flow of 0.09 m, 0.1025 m and 0.1241 m, vane height of 0.02 m, 0.04 m, and 0.06 m and vane rows of 0, 1, 2, 3, and 4 was studied. It was seen that in the case of four vane rows, generation of circulation field was large and was extended to a greater distance; hence the mixing was highest in the case of four arrays of vanes. The number of arrays is proportional to the transverse mixing of the pollutants. Variation of ratio of transverse mixing coefficients with vane and without vane with ratio of height of vane and depth of flow has been studied. It is found that as the vane size increases there is a drastic increase in the transverse mixing coefficient. This is due to the fact that a high magnitude of transverse circulation is generated in the flow as vane size increases. A predictor to estimate transverse mixing in the presence of submerged vane rows was developed and it vi was observed that as the depth of flow was increased the transverse mixing was reduced due to increase in submergence over the vanes. Transverse mixing was observed to be proportional to the height of vane and number of vane rows. Further, the transverse mixing coefficient for higher arrays/rows of vanes is high. For an example for depth of flow of 0.1241 m, the transverse mixing coefficient for four, three, two and one array of vanes are 23, 17, 13, and 7.5 times, respectively higher than the transverse mixing coefficients with no vane condition. However, such order of increase in the transverse mixing coefficient with vane for lower depth of flow is low. Flow with vane height/depth of flow of the order of 0.25 is not significant for enhancement in the transverse mixing. Transverse mixing length is an important parameter in the establishment of longitudinal movement of pollutants because it is assumed that after transverse mixing is complete then only the motion in longitudinal will start prominently (Fischer, 1979; Rutherford, 1994). Transverse mixing length was calculated for 98% mixing case and was observed that for 0.06 m vane and depth of flow = 0.1241 m, tracer mixes 11 times faster than its absence. In case of 0.04 m vane and depth of flow = 0.1025 m, this mixing length in presence of vane was around 3.5 times shorter than what has to be without vane. For 0.02 m vane and depth of flow = 0.09 m, tracer was observed to mix 2 times faster than without vanes. This study indicates that submerged vanes can be used for the enhancement of the transverse mixing subjected that morphological changes in the alluvial streams are not noticeable.